Anthropic's Claude Tracker Accused of Spying; Engineer Confirms Experiment Over

Anthropic, the artificial‑intelligence firm behind the Claude chatbot, faced fresh criticism after internal documents revealed that the company had been collecting detailed user interaction data without explicit consent. The revelations, first reported by a former Anthropic engineer, described a covert “experiment” in which the firm logged prompts, responses, and metadata from millions of users to refine its language models. Critics have labeled the practice a breach of privacy, prompting calls for regulatory scrutiny and heightened transparency from AI developers.

The engineer, who requested anonymity, confirmed that the data‑gathering initiative was concluded last month and that Anthropic had halted the program following internal review. In a brief statement, Anthropic said the experiment was intended to improve safety and performance but acknowledged that the approach did not meet the company’s own standards for user consent. The firm announced it would implement stricter data‑handling policies, provide clearer disclosures, and cooperate with any investigations by authorities.

The episode adds to growing concerns about how AI companies handle user information and underscores the sector’s need for clearer privacy frameworks. As regulators worldwide consider tighter rules on AI data practices, Anthropic’s recent actions may serve as a benchmark for industry‑wide standards on user consent and transparency.